In this issue of the JCI, two reports (10, 11) define how distinct mechanisms, induced by RT, restrain STING signaling and can be blocked to amplify the immune-engaging and antitumor effects of RT.

The RNA N6-methyladenosine binding protein YTHDF1 suppresses tumor antigen cross-presentation on DCs by stabilizing lysosomal cathepsin encoding mRNAs (12). Wen, Wang, and colleagues (10) discovered that YTHDF1 was induced in peripheral blood DCs and associated with shorter progression-free survival of patients with non–small cell lung cancer after RT. The induction of YTHDF1 was dependent on IFN-I signaling and mediated by STAT2, downstream of the IFN α/β receptor (IFNAR). Ythdf1 deletion enhanced IFN-I production in DCs, and DC-specific deletion of Ifnar negated the antitumor effects of Ythdf1 deletion after RT. It was previously demonstrated that YTHDF1 stabilizes lysosomal cathepsins (12) and that STING signaling occurs in vesicles that are degraded by lysosomes after activation (13, 14). Wen et al. reported several effects of RT: cathepsins were induced along with YTHDF1, STING physically interacted with cathepsins, and STING signaling and associated IFN responses increased after cathepsin inhibition in vitro and in vivo (10). Importantly, two strategies improved tumor control: in vivo cathepsin inhibition combined with RT, as well as Ythdf1 deletion or inhibition in tumor antigen–loaded DC vaccines delivered to mice receiving RT (10). Thus, YTHDF1 mediates immunological resistance to radiation, and targeting YTHDF1 or cathepsins to prolong STING signaling after RT may represent a route to improve patient outcomes.

Zhang, Deng, Wu and colleagues (11) sought to discover genes that determine IFN induction after RT using a CRISPR-KO screen targeting metabolic genes in a nasopharyngeal cancer cell line. Deletion of the antioxidant hemeoxygenase 1 (HO-1) potently promoted IFN-β production after RT. Inducible knockdown of HO-1 in various murine models sensitized tumors to RT. Using RNA silencing and analysis of cell signaling via immunoblotting, the authors determined that HO-1 ablation promoted IFN-I via cGAS/STING signaling. Intriguingly, HO-1 ablation increased cGAMP production after RT, but also enhanced IFN-β production after cGAMP treatment, implying that HO-1 acts on cGAS/STING signaling at multiple nodes. Expression of HO-1 has previously been shown to be promoted by ROS, and HO-1 is believed to mediate antiinflammatory effects primarily through the action of its enzymatic products (15). However, RT and IFN-β induced HO-1 expression independent of ROS, and HO-1 suppressed IFN-I induction after RT independent of its catalytic activity. Induction of full-length and cleaved versions of HO-1 were observed after RT; with cleavage of HO-1 only being observed after RT (11). Using truncated and mutant HO-1 variants, the authors discovered that full-length HO-1 remained anchored to the ER via its transmembrane domain, while RT-induced HO-1 cleavage disrupted the transmembrane domain of HO-1, causing its nuclear relocalization. Nuclear (truncated) HO-1 directly bound to cGAS to prevent its nuclear export after RT, thereby limiting the production of STING-activating cGAMP in the cytosol. Independent of this function, full-length (uncleaved) HO-1 was retained at the ER, where it was shown to directly interact with STING, prevent its oligomerization, and preclude its association with TBK1, the kinase responsible for IRF3 phosphorylation and subsequent IFN production after STING activation. A screen for inhibitors of HO-1 that potentiated RT-induced IFN-β production yielded a candidate (HO-1–IN-1) that promoted cGAMP production and STING activation. This inhibitor promoted antitumor and immunogenic effects of RT in vivo, demonstrating preclinical feasibility of targeting HO-1 to augment STING-mediated inflammation after RT. HO-1 expression was associated with poor outcomes for patients with nasopharyngeal cancer and shorter survival in patients with brain cancer treated with RT. This work identifies HO-1 as a targetable mechanism that restrains STING signaling after RT and, more broadly, implicates HO-1 as a regulator of STING signaling.

Both studies (10, 11) identify compelling mechanisms, supported by murine models and associated human correlatives, that impede RT-induced STING signaling and, thus, antitumor T cell immunity. The authors of each study demonstrated proof of principle that STING signaling can be augmented to boost in situ vaccination effects of RT: either by blocking cathepsin-mediated degradation of active STING signaling complexes in DCs (10), or by blocking the suppressive function of HO-1 on both cGAS and STING in cancer cells (11) (Figure 1).

STING suppression after RT limits antitumor T cell immunity. Zhang et al. (11) demonstrated that RT induces HO-1 expression and cleavage in cancer cells. This process leads to two mechanisms by which HO-1 suppresses cGAS/STING signaling. First, cleaved HO-1 localizes to the nucleus and binds to cGAS to prevent its nuclear export — preventing cGAS-dependent production of cytosolic cyclic dinucleotides (CDNs), such as cGAMP, that activate STING. Second, uncleaved HO-1, which retains its transmembrane domain, remains at the ER and directly interacts with STING to prevent its oligomerization, ER lumen curvature, and interaction with TBK1, impeding downstream signaling from STING. Together, these effects reduce the amount of intracellular CDN production and IFN-I secretion, limiting the delivery of these key immunostimulatory molecules to DCs and other cells within the tumor microenvironment. Wen et al. (10) discovered that RT induces the expression of YTHDF1 through IFN-I:IFNAR–dependent STAT2 activation, which directly binds to the Ythdf1 promoter to promote YTHDF1 transcription in DCs. YTHDF1 then binds to cathepsin mRNA to support its translation, leading to an overall increase in cathepsin expression and presence in lysosomes. STING activation, elicited by either engulfed cancer cell DNA recognized by cGAS or through import of extracellular CDNs, leads to the oligomerization of STING and production of vesicles from which productive STING signaling occurs. These vesicles are degraded by cathepsins in lysosomes, attenuating STING signaling. Ultimately, cancer cell and DC-intrinsic STING signaling can induce DC maturation/activation that leads to priming of tumor antigen–specific CD8+ T cells by DCs that recognize and kill cancer cells. Both HO-1 and YTHDF1 impair DC maturation, cross-presentation, and/or antitumor T cell immunity after radiation.